HPE ProLiant Servers

Both are current and both are the right buy depending on the model. HPE introduced Gen12 in February 2025 with Intel Xeon 6 across DL320, DL360, DL380, DL380a, DL580, and ML350, then added AMD EPYC Gen12 models (DL325, DL345) in mid-2025. But DL385 (2-socket AMD flagship), DL560 (4-socket Intel), and the entry towers ML30, ML110, and MicroServer remain Gen11 with no Gen12 successor announced yet, so Gen11 is the current, correct answer in those segments, not a legacy fallback.

iLO 7 adds a dedicated secure enclave processor and is the first server management controller HPE has shipped with quantum-computing-resistant readiness, built to meet FIPS 140-3 Level 3 requirements using hash-based cryptography. iLO 7 ships only on Gen12 models; DL385 Gen11, DL560 Gen11, and the Gen11 towers still run iLO 6, which remains fully supported and Silicon Root of Trust protected.

The DL360 Gen12 is a 1U, two-socket server best when rack density matters and you need a compact node with up to 20 EDSFF drives. The DL380 Gen12 is a 2U server with room for up to 36 EDSFF drives, more GPU slots, and rear storage options, making it the more flexible default for mixed enterprise and hybrid cloud workloads. Both run the same Intel Xeon 6 processors up to 144 cores.

AMD EPYC ProLiant models deliver the highest core counts per socket for dense virtualization and consolidation. The single-socket DL325 Gen12 and DL345 Gen12 both reach up to 192 cores on 5th Gen AMD EPYC with up to 6 TB of DDR5, while the two-socket DL385 Gen11 scales to 160 cores and has repeatedly placed first in MLPerf Inference and TPC-H price-performance benchmarks. Intel Xeon 6 models can be the better fit for per-core-licensed software or specific accelerator-tuned workloads.

The DL380a Gen12 is a 4U, dual-socket server purpose-built for AI training and inference, supporting up to 10 double-wide GPUs including NVIDIA H100 NVL, H200, L40S, and RTX PRO 6000 Blackwell Server Edition, with three dedicated GPU power domains and direct liquid cooling. It claimed eight number-one positions in MLPerf Inference v6.0 llama-focused tests, making it the AI inferencing workhorse of the current ProLiant lineup.

Yes. The ML30 Gen11 and MicroServer Gen11 remain HPE's current entry towers for small offices and edge sites, both now offering Intel Xeon 6 or Xeon E-2400 options with optional NVIDIA RTX A400 GPU support. The ML350 Gen12 scales up as a two-socket, rack-convertible tower for growing mid-size sites needing up to 8 TB of DDR5 and 24 NVMe drives.

HPE Compute Ops Management is a cloud console that monitors, updates, and secures ProLiant servers regardless of generation, from data center to edge. It applies firmware baselines and configuration policies fleet-wide and is stated to cut maintenance effort by up to 75%, so a mixed Gen11/Gen12 estate still behaves like one managed platform.

Request pricing directly through Uniqcli's quote workflow at /quote, where you can specify the ProLiant model, processor, memory, and storage configuration along with any GPC direct, SAP, FAR-based orders, or GSA eBuy requirements. Uniqcli, as an authorized HPE partner, scopes the exact build and returns a configured quote rather than a generic price list.

Yes. Through HPE GreenLake you can consume ProLiant compute pay-per-use with capacity staged on site, paying for what you use while HPE manages buffer capacity. Uniqcli can model whether a capital purchase or GreenLake consumption fits your budget cycle and growth plans.